Security Education for the new Generation

Transcription

1 Security Education for the new Generation SESSION SESSION ID: ID: MASH-W02 Wednesday, Feb 26, 9:20 WEST 3018 Jacob West Chief Technology Officer HP Enterprise Security Products Matt Bishop Professor University of California, Davis

2 A Message from Matt Sorry I cannot be here! Jacob and I prepared the slides together, and have known each other for years. This talk represents our views. 2

3 Defining security State of play today Opportunities Existing resources 3 Conclusion

4 Security Stuff vs. Secure Stuff Security Stuff Responsible for security Focus on security activities Opportunities for apprenticeship Clear career path for the motivated Doing Stuff Securely Responsible for stuff Impact on security, but not focus Few opportunities to learn Hard to map interest to career

5 What Are We Talking About? Robust programming Programming that prevents abnormal termination or unexpected actions A secure program conforms to a security policy And implicitly requires robustness, but robust programming does not require such conformance Here, secure is used in the sense of robust

6 Why This Matters Example: Exploit a buffer overflow to force a program to do things it should not Definitely non-robust Does not handle invalid input properly Is it non-secure? That depends if one can Use the buffer overflow to do things that the security policy disallows (yes) Use the buffer overflow to do only things that I could already do (no)

7 Defining security State of play today Opportunities Existing resources 7 Conclusion

8 Problem: 84% of Breaches Target Software Networks Hardware Applications Security Measures Intellectual Property Switch/Router security Firewalls Customer NIPS/NIDS Data VPN Net-Forensics Anti-Virus/Anti-Spam Business DLP Processes Host FW Host IPS/IDS Trade Vuln. Assessment Secrets tools

9 Software Security Today The line between secure and insecure is often subtle Many seemingly non-security decisions impact security Small problems hurt a lot A single bad line of code can put a company in the news Smart people make dumb mistakes As a group, programmers repeat the same security mistakes over and over We need non-experts to get security right Security and development are both full time jobs

10 Becoming a Doctor years

11 Becoming a Programmer Enter Workforce 4 years Learn Everything On-the-Job Training? Certification?

12 Top 9 Undergraduate Computer Science Programs Required classes with security content Restricted electives with security content Restricted electives with security focus Berkeley Carnegie Mellon Cornell Georgia Tech Michigan MIT Stanford Illinois Austin

13 Top 9: Courses with Security as a Focus Junior/Seniors specialize Only 3 of 9 offer security track Cornell: Security & Trustworthy Systems Track (4 classes) Michigan: Security Track (4 classes) Austin: Information Security Certificate Program (5 classes) Berkeley Carnegie Mellon Graduate Cornell Georgia Michigan MIT Stanford Illinios Austin Undergraduate

14 Top 9: Courses Focus on Traditional Security Software security appears at 3 of 9 General Cryptography Network Web Forensics Software Mobile Privacy Policy/Risk

15 The Fundamental Problem We don t write software that is robust Some exceptions in special cases We don t build systems to meet security requirements Many different models for developing software

16 What Will Drive Improvement? Commercial Financial savings (avoid cleaning up messes) Simpler maintenance Improved reputation Government Financial savings (avoid cleaning up messes) Simpler maintenance National security Software liability?

17 Software Liability You can t say I m not responsible for anything Chain of distribution (e.g. supply chain) liability exists now You can limit liability somewhat by defining use and environment Then you re liable in that context but (probably) not in others It is coming... EULAs may not be enforceable (adhesion contracts)

18 So What s Holding Us Back? Commercial and Government Need to spend more money Longer time-to-market No legal liability for bad software Need to pay more attention to installation, maintenance, and use Lack of people to write good code

19 So What s Holding Us Back? Academia Robust coding not seen as integral to programming Textbooks loaded with examples of non-robust programming Lack of support for enforcing and grading for robust coding Ties into lack of graders who really know about this Lack of faculty who understand robust programming And intimidation factor for those who know they don t understand it

20 Lack of Resources Assurance costs! Industry expected to deliver secure, robust products without resources for the extra effort required to deliver them Academia expected to teach and reinforce robust programming without resources for the extra effort in supporting this

21 Lack of People Need to teach people how to write robust programming Need to emphasize the practice, both in education and industry Continuous practice is key to reinforcing, maintaining, extending skills

22 Defining security State of play today Opportunities Existing resources 27 Conclusion

23 Focus for Rest of Talk Everyone lacks resources! How can industry and government work with academia? Carrots, not sticks Security tuned to environment and use What is secure varies among companies and government organizations

24 What Do We Need to Teach Secure Programming? Summit on Education in Secure Software (SESS) Diana Burley The George Washington University Matt Bishop University of California at Davis Funded by NSF 60 participants: academia / industry / government nob.cs.ucdavis.edu/~bishop/notes/2011-sess/2011-sess.pdf

25 SESS Objectives Engage stakeholders from academia, industry, and government to discuss teaching secure programming Use discussion as basis for a collaborative effort to develop a comprehensive agenda for secure software education

26 Recommendations Increase faculty who understand the importance of secure coding Establish professional development opportunities for faculty Integrate computer security content into existing courses Provide faculty support for the inclusion of security content Require at least one computer security course for all college students

27 Recommendations Promote collaborative problem solving and solution sharing Encourage partnerships and collaborative curriculum development Use innovative teaching methods to strengthen the foundation of computer security knowledge Develop metrics to assess progress toward meeting educational goals Highlight the role that computer security professionals should play in key business decision making processes

28 Summary Holistic view of secure education suggests programmers and nonprogrammers alike must be educated in the core principles Structural enablers Cultural shift among faculty and industry stakeholders that supports the development of a holistic view of software security Identification of measurable objectives and corresponding measurements Development of national licensure programs Alignment of expectations for university education and realistic

29 What Can Academia Do? Include robustness in evaluation of programs and projects Create a secure programming clinic Like an English clinic, or a writing clinic for law schools Provide supplementary material for textbooks, classes These should emphasize robust programming

30 What Can Industry Do? Key is to show more than say secure development is important Make clear that the skills are important for hiring Mention their need in job openings Preference to those with skill in this also helps

31 Work With Students and Faculty Internships Students love these; good recruiting tool Tasks requiring robust programming emphasize its importance to students Help teach students Review students code Team with colleges in senior/capstone projects

32 What Will This Do? Increase student demand If students see it as important, they will ask about it in class, evaluate programs, faculty in part on it Increase your visibility Good recruiting tools A corporate good citizen

33 Government Support Act like an industry Government can also fund programs (e.g. DHS and the Software Assurance Curriculum Project) Programs should support future faculty (as well as engineers) People willing to commit to teaching Imperative: target funding towards this specific purpose Require funding to be used for supporting robust programming If done as adjunct, likely to disappear in the main purpose of the funding

34 Defining security State of play today Opportunities Existing resources 42 Conclusion

35 Existing Resources Industry academic enablement programs Tradeshows and conferences Competitions and contests Training and certification Specialized university programs

36 Industry Academic Enablement Programs Focused on facilitating university education in security Notable Programs Cisco IBM Delivery Methods Hewlett-Packard Microsoft Direct to students or young professionals with certification programs Define a methodology, enlist industry specialists to assist with delivery Collaborate with universities and non-profits on research

37 Tradeshows and Conferences Notable Conferences IEEE Symposium on Security and Privacy ACM Computer and Communications Security The Colloquium (Colloquium on Info. System Security Education) USENIX Security Symposium Goals Share advancements in research Enhance with curricula security-centric topics

38 Competitions and Contests Notable Competitions CyberPatriot National Collegiate Cyber Defense Competition UCSB ictf Cyber Security Awareness Week (CSAW) DEFCON Why Capture-the-Flag? Goal-oriented and rewards both participation and success Opportunity to network with peers and industry professionals

39 Training and Certification Notable Programs (ISC) 2 SANS Institute CompTIA Motivations and Objectives Industry effort to develop and ensure baseline skillsets Differentiate candidates for human resources and hiring managers Validate and recertify relevant security experience

40 Scholarship Programs Government Scholarships (for work commitment) CyberCorp: Scholarships for Service DOD Information Assurance Scholarship Program Private Scholarships (ISC) 2 Scholarships (Community College, Undergrad, Grad) Armed Forces Communications and Electronics Association (Community College, Undergrad, Grad) National Security Scholars Program (Undergrad) Symantec Graduate Fellowship (Grad) Applied Computer Security Associates (Undergrad, Grad)

41 Scholarship for Women Studying Information Security Support women with a demonstrated interest in security, through coursework, internships or work experience to complete a Bachelors or Masters degree Must be entering junior or senior year of Bachelors or first year of Masters Administered by Applied Computer Security Associates (ACSA) and Committee on the Status of Women in Computer Research (CRA-W) Awarded a single $10k scholarship annually pre-2014 Includes attendance at ACSA, CRAW, as well as internship opportunities

42 Defining security State of play today Opportunities Existing resources 71 Conclusion

43 Many Myths Myths about universities confuse how academia can put teaching secure programming into practice There is no room If students learn to write secure programs Academic institutions are We know what to do And some questionable ideas don t help either Testing students knowledge Unsupported mandates

44 Myth #1 There is no room in curricula for a course on secure programming You don t need a separate course Simply check programs submitted during all courses for robustness Make writing robust programs beneficial (through grades or other mechanisms) Provide resources so students can see how to do this, or get help to do it

45 Myth #2 If students learn to write secure programs, the state of software and system security will dramatically improve Will companies accept increased cost, time to market? Will customers pay higher prices, endure longer development times? Will students be encouraged (required) to practice what they learned?

46 Myth #3 Academic institutions are hierarchical in organization Implication is that deans, provosts, presidents can order this taught Learning styles, environments differ May not be a 'best' or 'right' way to teach this

47 Myth #4 We know what to do and how to do it We have ideas, but don t know Needed: research on education Needed: funding, people to do this The Summit on Education in Secure Software suggested ways to do this SESS results are general Approaches must be tailored to various environments

48 Questionable Idea #1 Testing students knowledge Who creates the tests? Who is being tested? How do you know that you are testing what is important? (that is, the right thing ) Who determines what is an acceptable result? Teaching to the test, rather than to learn the material

49 Questionable Idea #2 Unsupported mandates The support has to come from somewhere It s like a zero-sum game What do you want to weaken? If you only have so many resources, something will have to give You don t want to weaken the core foundation of understanding why certain programming paradigms are critical

50 Conclusion We must all hang together, or we shall all hang separately. - Benjamin Franklin The state of practice can, and must, change Understand that academia is a different environment completely Teaching robust programming, and nothing more, will not help The marketplace must also change, as must current practice The public will be the main driver (unfortunately, probably with lawsuits)

51 Questions? SESSION ID: Jacob West Chief Technology Officer HP Enterprise Security Products Matt Bishop Professor University of California, Davis